项目名称: 基于协同语义计算的社交媒体信息扩散与可信性研究
项目编号: No.61202140
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 计算机科学学科
项目作者: 刘知远
作者单位: 清华大学
项目金额: 24万元
中文摘要: 以社交网络服务和微博为代表的社交媒体,是人们分享与交流信息的新平台。随着社交媒体的兴起,信息扩散模式发生深刻变化;信息源激增也容易诱发不实信息泛滥,给公共安全带来隐患。因此,社交媒体信息扩散机制和可信性感知问题成为社会关注焦点,是社会计算的前沿科学问题。针对现有研究手段囿于社交网络结构等表层分析的局限,本项目根据用户协同产生的海量内容,对用户、信息的语义属性及其复杂语义联系进行建模。基于对用户和信息的协同语义计算,进一步开展以下研究:研究适用于多用户、多信息复杂情形的社会影响力分析方法;研究信息多通道扩散的分析方法;综合机器智能与群体智能的信息可信性分析方法;研究跨社交媒体的信息扩散、社会影响力与信息可信性的统一分析与预测系统,在典型平台上验证有效性。本项目预期成果将深化社交媒体信息扩散机制与可信性感知的研究,对互联网规模社交媒体的信息组织与管理以及社交媒体时代的中文信息处理均深具意义。
中文关键词: 社会媒体;协同语义计算;信息扩散;信息可信性;表示学习
英文摘要: Social media, such as social network services (SNS) and microblogs, provides a novel platform for people to share and communicate information. With the rise of social media, the patterns of information diffusion profoundly change. Moreover, the dramatic increase of information sources will lead to misinformation overflow, a hidden danger to public interest. Information diffusion mechanism and information credibility thus turn into social focus, and also the cutting-edge science of social computing. Most existing research works are confined to surface analysis of the social network structure. To address the problem, this project will investigate the semantic models of users, information and their complex semantic relations according to the collaborative behaviors of social media users. Based on the collaborative semantics of users and information, this project will further carry out the following studies: (1) explore analysis methods of social influence that are feasible for the complex situation with multiple users and multiple information; (2) explore analysis methods of information diffusion through multiple channels among users; (3) explore analysis methods of information credibility by incorporating machine intelligence with collective intelligence in social media; (4) implement a unified cross-platform anal
英文关键词: social media;collaborative semantic computing;information diffusion;information credibility;representation learning